Disability from stroke is the major cause of neurological morbidity in the United States. The mechanisms responsible for recovery of neurological deficits due to cerebral infarction of hemorrhage are poorly understood. An improved understanding of the process would lead to more rational design of strategies to improve neurological recovery. We propose to investigate the role of basal ganglia and cerebellum in the process of neurological recovery from motor deficits following hemispheric stroke. The specific hypothesis is: Following damage to cortico-spinal neurons by hemispheric stroke (infarction or hemorrhage), recovery of neurological deficits in voluntary motor activity is mediated by modulation of the activity of subcortical motor systems, specifically by an increase in task related neuronal activity of ipsilateral globus pallidus and contralateral dentate nucleus. Therefore, performance of a recovered motor task that was initially impossible following hemispheric stroke will produce a greater task-related increase in regional cerebral blood flow (as a measure of local neuronal activity) in ipsilateral globus pallidus and contralateral dentate nucleus than the same task performed by the undamaged contralateral motor system. The specific aim is: To measure changes in regional cerebral blood flow (rCBF) with positron emission tomography (PET) in globus pallidus and dentate nucleus bilaterally during best possible performance of volitional motor task paradigms in patients 1 - 2 weeks and 6 months after hemispheric stroke (hemorrhage or infarction). In patients who show neurological recovery, the magnitude of the rCBF increases (deltaCBF) in globus pallidus and cerebellum of the damaged motor system will be compared to deltaCBF produced by the same task performed by the undamaged side. To confirm this hypothesis we would need to find (1) at 6 months, significantly greater deltaCBF in these structures as compared to non- damaged side for the same task; (2) significantly greater deltaCBF on the damaged side at 6 months for a recovered task than at 7 days. These two changes would also have to be restricted to those patients who improve or correlated with the degree of improvement.